Coefficient of variation

As explained in Demand Forecasting Best Practices

The coefficient of variation is the ratio of a quantity's standard deviation to its mean, often used to express relative variability.

When planners ask whether a product is easy to forecast, they may look at the coefficient of variation (COV), the ratio of standard deviation to mean: COV = σ/μ. Often expressed as a percentage, it measures variation around the historical mean—not unexpected future variation. section 10.2

COV can therefore mislead forecasting decisions. Trend and seasonality can produce high COV even when a suitable forecasting method predicts the product well. Historical COV also benefits from data leakage because it compares observations with a mean calculated using information from the same period. section 10.2.1 section 10.2.2

Treat COV as a descriptive variability measure, not a default forecastability score:

• Use it cautiously for flat, patternless demand, where it can provide a relatively good indication of overall forecastability. section 10.2.1
• Do not use it to rank products on an XYZ forecastability axis when demand contains trend or seasonality. section 13.2.2
• Assess forecastability with historical forecast errors against simple benchmarks such as moving averages. section 10.3 section 13.2.2

Why the same metric can misclassify products

A stable product provides a relatively favorable COV comparison with a moving-average benchmark. By contrast, products with an upward trend or seasonal pattern can show high COV while remaining easy to forecast with a moving average or seasonal moving average. section 10.2.1 section 10.2.2

Figure 10.6 COV vs. moving average for a product with trend

Compare COV with moving-average performance for a product with trend, where high historical variation does not establish poor forecastability. section 10.2.2

Figure 10.7 COV vs. seasonal moving average for a seasonal product

Inspect the seasonal example to see why COV can flag a structured, forecastable demand pattern as difficult. section 10.2.2

How Demand Forecasting Best Practices covers it

How deep it goes

Defines COV as σ/μ, tests flat, trending, and seasonal demand, and explains data leakage.

Examples

Compares stable, trending, and seasonal products with moving-average forecasting benchmarks.

What you can do after

Evaluate forecastability by comparing forecast errors with simple moving-average benchmarks rather than COV.

How it connects

Connects COV with forecastability, forecast error, moving averages, trend, seasonality, and data leakage.

liveBookmanning.com

Read Demand Forecasting Best Practices on liveBook Get the book at manning.com

Related concepts

• meanuses · A mean is a measure of central tendency, calculated as the arithmetic average of values or, for a random variable, as the probability-weighted average of its possible values.
• standard deviationuses · Standard deviation is the nonnegative square root of variance and measures the typical dispersion of values around their mean in the original units.
• forecastabilitycontrasts with · Forecastability is the degree to which a demand series can be predicted accurately from available historical information.
• auxiliary lossused by · An auxiliary loss is an additional training objective or loss term optimized alongside a primary objective to encourage a desired property or improve learning.
• seasonality effectcontrasts with · A seasonality effect is a recurring pattern of variation in data associated with regular temporal, calendar, or time-of-year cycles.
• ABC XYZ analysiscontrasts with · ABC XYZ analysis is a two-dimensional segmentation method that assigns items to categories using an ABC dimension and an XYZ dimension.

This explainer was generated by Manning's knowledge graph from the text of Demand Forecasting Best Practices. The section links point to the parts of the book it draws on.